CN114998165A - Histogram transformation based uniform exposure image contrast enhancement method - Google Patents
Histogram transformation based uniform exposure image contrast enhancement method Download PDFInfo
- Publication number
- CN114998165A CN114998165A CN202210664585.1A CN202210664585A CN114998165A CN 114998165 A CN114998165 A CN 114998165A CN 202210664585 A CN202210664585 A CN 202210664585A CN 114998165 A CN114998165 A CN 114998165A
- Authority
- CN
- China
- Prior art keywords
- histogram
- image
- sub
- threshold value
- histograms
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000009466 transformation Effects 0.000 title claims abstract description 22
- 230000006870 function Effects 0.000 claims abstract description 18
- 238000012546 transfer Methods 0.000 claims abstract description 12
- 238000005315 distribution function Methods 0.000 claims description 7
- 230000001186 cumulative effect Effects 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 2
- 238000010008 shearing Methods 0.000 abstract 2
- 238000012545 processing Methods 0.000 description 10
- 230000016776 visual perception Effects 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 3
- 230000015556 catabolic process Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 210000004072 lung Anatomy 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004321 preservation Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000005481 NMR spectroscopy Methods 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000002591 computed tomography Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20132—Image cropping
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a histogram transformation-based uniform exposure image contrast enhancement method, which transforms a histogram through a histogram probability density function, a two-pole boundary threshold method and gamma transformation. The method comprises the steps of firstly, calculating a shearing threshold value by using a two-pole boundary threshold value method, utilizing the shearing threshold value to shear an original histogram, then respectively calculating two-pole boundary sub-histogram transfer functions, reallocating gray levels contained in the two-pole boundary sub-histogram, correcting the distribution of the original histogram, and performing two-polarization gamma self-adaptive stretching on an image according to the corrected histogram so as to obtain an image with enhanced contrast. The method effectively solves the problem that the existing algorithm excessively enhances the uniform exposure image, reasonably improves the image contrast, and keeps the integrity of information such as natural appearance display, texture detail, edge characteristics and the like of the image.
Description
Technical Field
The invention belongs to the field of image enhancement, and particularly relates to a uniform exposure image contrast enhancement method based on histogram transformation.
Background
During image acquisition, due to factors of the image acquisition equipment (such as improper parameter adjustment and inherent limiting properties of the equipment) or conditions of the object (i.e. different absorption and reflection properties of the light source), a complex noise model exists in signals during acquisition and transmission, so that some uniformly exposed images have low-contrast characteristics, which generally causes great obstruction to subsequent application of such uniformly exposed images. In contrast to uniformly exposed images, are non-uniformly exposed images that exhibit a very bright or very dark visual perception due to overexposure or underexposure.
Common low-contrast uniform exposure images include computed tomography images, nuclear magnetic resonance images, and natural images acquired under a weak light source, among others. The difficulty in distinguishing the details, textures, edges and other regions of interest of the low-contrast uniformly-exposed image causes great trouble to image users, so that an image enhancement algorithm for the low-contrast uniformly-exposed image is needed to be invented. Improving the image characteristics of digital images clearly reveals that perceptual information is the main goal to improve image contrast. In various image processing applications, image enhancement is used as a preprocessing stage, such as low-light-level image enhancement, dynamic range adjustment, automatic contrast enhancement, and remote sensing image enhancement. Furthermore, image contrast enhancement plays a crucial role in machine learning.
The method of performing contrast transformation using a histogram is called histogram transformation, and generally refers to processing the distribution of a target histogram, which is a method of effectively enhancing the contrast of an image. Common histogram transformation methods are histogram equalization, normal distribution, general transformation, and parametric transformation, etc., wherein histogram equalization is a well-known method, but its over-enhancement result on medical image processing is often unacceptable to clinicians in some cases. In recent years, there is a class of image contrast enhancement algorithms that modify histograms based on plateau limits, using histogram plateau limits to cut the histogram, with the remaining pixels being reassigned to the relative vacancies in the histogram, using the modified histogram to compute the transfer function to enhance the contrast of the image. Such algorithms have been effective in improving the contrast of uniformly exposed images, but in a few cases, can disrupt the natural display of image features.
The detail information of the uniformly exposed image is rich, and the current contrast enhancement algorithm often causes the excessive enhancement of the image, thereby destroying the natural display and unique appearance of the image and causing the loss of the detail information of the interested area of a user. Therefore, the proposal and application of contrast enhancement algorithm should follow the principles of enhanced image preserving natural appearance, attractive contrast, less degradation, reasonable detail preservation, etc. Therefore, on the premise of protecting the natural display of information such as image details, textures, edges and the like, the problem of enhancing the image contrast still exists. Therefore, it is desirable to provide an image contrast enhancement algorithm that can maintain the overall details of the image, complete the texture information, and improve the visual effect of contrast.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the problem that the image is excessively enhanced by the existing algorithm is solved, the contrast of the uniformly exposed image is reasonably improved, the integrity of information such as natural appearance display, texture details, edge features and the like of the image is kept, the influence of the image processing algorithm on the self-feature display of the image is reduced, and a good foundation is laid for further use of the image.
The technical scheme adopted by the invention is as follows: a histogram transformation-based uniform exposure image contrast enhancement method comprises the following steps:
step 1), firstly, calculating a gray level histogram of an image;
step 2), calculating a gray level intensity value boundary threshold value of the image according to a new two-stage boundary threshold value formula, namely a dark part threshold value and a bright part threshold value;
step 3), after two gray level thresholds are obtained, the histogram in the step 1) is cut by using the thresholds to obtain three sub-histograms;
step 4), respectively calculating the probability density function and the cumulative distribution function of the two boundary sub-histograms;
step 5), respectively calculating transfer functions of the two boundary sub-histograms mapped to the middle sub-histogram by utilizing the cumulative distribution function;
step 6), redistributing the pixels of the two boundary sub-histograms to a middle sub-histogram region according to the two transfer functions to obtain an in-process image and a corresponding corrected sub-histogram;
and 7), expanding the corrected sub-histograms to the global state, namely performing dual-polarization gamma stretching on the histograms of the images in the process to obtain the final enhancement result.
Further, the method corrects the original histogram distribution based on histogram transformation, redistributes two-pole boundary pixels, and performs two-polarization adaptive gamma transformation stretching on the image according to the corrected histogram so as to obtain the image with enhanced contrast.
The principle of the invention is as follows: a histogram transformation based contrast enhancement method for a uniform exposure image is disclosed, which combines a histogram probability density function, a new histogram two-pole boundary threshold value method and gamma transformation, calculates the histogram by using the two-pole threshold value method to obtain two boundary threshold values, cuts the histogram by using the boundary threshold values, calculates a transfer function by using the probability density function and an accumulative distribution function, redistributes the cut histogram to an uncut histogram area to obtain a modified histogram, and finally performs adaptive gamma transformation on the modified sub-histogram to perform histogram stretching, thereby obtaining a final contrast enhancement image.
Compared with the prior art, the invention has the following advantages:
the invention provides a histogram transformation-based contrast enhancement method for a uniformly-exposed image. Compared with the histogram equalization in the traditional technology, the method can not damage the natural appearance of the image, can save reasonable details and clear characteristics, solves the problem of over-enhancement of the histogram equalization, is obviously more natural and better accords with the visual perception of human eyes. Compared with the contrast enhancement algorithm based on the platform limit correction histogram, which is proposed in the last two years, the enhanced image result has more attractive contrast and has more advantages in the aspect of protecting the whole visual effect of a uniform exposure image.
Drawings
Fig. 1 is a flow chart of the algorithm technical scheme of the invention.
FIG. 2 is a diagram of dividing and clipping a histogram in an embodiment.
Fig. 3 is a comparison diagram before and after histogram modification in the embodiment, where in fig. 3(a) is hr (l) and fig. 3(b) is hm (l).
Fig. 4 shows the actual implementation result of the algorithm parameter λ being 0.07 in the embodiment, where fig. 4(a) is the original image of the lung CT in the public data set LIDC and the corresponding histogram distribution, fig. 4(b) is the result image of the present invention and the corresponding histogram distribution, and fig. 4(c) is the comparison algorithm processing image and the corresponding histogram distribution.
Fig. 5 shows the actual implementation result of the algorithm parameter λ of 0.07 in the example, in which fig. 5(a) is the original image of the photo man in MATLAB database and the corresponding histogram distribution, fig. 5(b) is the image of the result of the present invention and the corresponding histogram distribution, and fig. 5(c) is the comparison algorithm processed image and the corresponding histogram distribution.
Fig. 6 shows the actual implementation result of the algorithm parameter λ being 0.07 in the embodiment, where fig. 6(a) is the original image of the lung CT in the public data set LIDC and the corresponding histogram distribution, fig. 6(b) is the result image of the present invention and the corresponding histogram distribution, and fig. 6(c) is the comparison algorithm processing image and the corresponding histogram distribution.
FIG. 7 is a comparison of objective evaluation indexes of the image processing algorithm of the present invention and the existing algorithms.
Detailed Description
In order to make the purpose and technical solution of the present invention more apparent, the present invention will be further described in detail with reference to the following embodiments and the accompanying drawings.
Referring to fig. 1, the invention discloses a histogram transformation-based uniform exposure image contrast enhancement method, comprising the following steps:
1) inputting a digital image.
2) Calculating a gray level histogram of the image according to a formula: h (l) is n (l), n (l) represents the number of pixels of the image gray level l, and h (l) is the gray histogram statistics of the image.
3) According to the two-pole boundary threshold formula:
the intensity value (gray level) boundary threshold values of the image are calculated, respectively. Threshold represents the total number of pixels of the boundary, namely the pixels to be redistributed in the next step, and the value should keep the histogram distribution of the main dark part and the histogram distribution of the main bright part of the histogram, so that the detail information, the edge and the overall visual perception of the image can be kept to the maximum extent. Lmin represents a dark part boundary Threshold value, Lmax represents a bright part boundary Threshold value, gray values meeting Threshold are obtained by respectively calculating the number of boundary accumulated pixels, and then rounding operation is carried out on the Lmin and the Lmax. The value range of lambda is [0,0.5], ensuring that Lmin is less than Lmax.
4) The histogram h (l) is cut and divided by using two-level boundary threshold values Lmin and Lmax to obtain three sub-histograms as shown in fig. 2, where hr (l) is shown in fig. 2(C), hmin (l) is shown in fig. 2(B), and hmax (l) is shown in fig. 2 (D).
5) The cumulative distribution functions of the sub-histograms hmin (l) and hmax (l) are calculated, respectively: cdfmin, cdfmax; cumulative distribution functions are typically derived from histogram probability density functionsWhere, p (l) ═ h (l)/sum (h (l)) is the probability of the gray level l.
6) The transfer functions mapped to the two boundary sub-histograms hmin (l), hmax (l) to the sub-histogram hr (l) are calculated: TFmin ═ lmn +1) + Lmin × cdfmin, TFmax ═ Lmax +1) - (L-Lmax-1) × cdfmax. TFmin is the transfer function of the pixel from Hmin (l) to Hr (l), and TFmax is the transfer function of the pixel from Hmax (l) to Hr (l). Common methods such as histogram equalization map pixels of certain gray values robustly to boundaries, thus leading to either excessive enhancement of image processing results or degradation of the natural appearance display. This step 6) implements the redistribution of boundary pixels for the two transfer functions, which is innovative in that the distribution characteristics of the dark and light parts of the histogram are preserved to the maximum extent and all pixels are made to participate in the final histogram polar stretching.
7) The pixels of the two boundary sub-histograms are redistributed to the non-clipped histogram region according to the two transfer functions, so as to obtain a modified sub-histogram hm (l), as shown in fig. 3, where hr (l) is shown in fig. 3(a) and hm (l) is shown in fig. 3 (b). The border histogram of the two dotted line parts has a reassigned value of 0, and the mean value of hm (l) represented by the solid line part is increased correspondingly. This step 7) redistributes all pixels according to the main distribution characteristics of the original histogram, and can greatly preserve the appearance visual perception effect of the image according to the maximum entropy principle.
8) And performing adaptive gamma stretching on the corrected sub-histograms: y ═ x + eps) γ And the final enhancement result can be obtained. y represents the pixel output intensity value, x represents the pixel input intensity value, eps represents the base of the natural logarithm, is a mathematical constant, approximately equal to 2.718281828, and γ is a power of a function.
9) As shown in fig. 4, 5, and 6, the result was an example in which λ was 0.07. The result of the figure shows that the invention greatly improves the image contrast and simultaneously has complete information preservation of the image. FIG. 7 shows the objective evaluation index comparison of the present invention with the existing comparison algorithm, the image retains the information capacity after the information entropy evaluation processing, the larger the value is, the richer the image is; the larger the contrast is, the more gray levels the image is processed to occupy, and the richer the image information is, so that the pockets and the lines of the overcoat of the photographer can be seen on the image processed by the method in the invention in the picture 5; AMBE evaluates the ability of the proposed method to maintain the average luminance of the image, a small AMBE indicates that the average luminance of the enhanced image is close to or equal to the average luminance of the input image, and vice versa. PSNR is used to compare the noise level present in an image, and a high PSNR indicates that the existing noise is not amplified by the image processing technique. In conclusion, compared with other contrast stretching methods, the bi-polarization adaptive gamma stretching method disclosed by the invention processes more pixels and retains the characteristic distribution characteristics of the histograms of the dark part and the bright part, so that the entropy of the image information is closer to the original image, the contrast is improved more obviously, the detailed information is displayed more naturally, and the visual perception effect is better.
The above description is only an embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can understand that the modifications or substitutions within the technical scope of the present invention are included in the scope of the present invention.
Claims (2)
1. A histogram transformation-based uniform exposure image contrast enhancement method is characterized in that: the method is realized by the following steps:
step 1), firstly, calculating a gray level histogram of an image;
step 2), calculating a gray level intensity value boundary threshold value of the image according to a two-pole boundary threshold value formula, wherein the gray level intensity value boundary threshold value can be called as a dark part threshold value and a bright part threshold value;
step 3), after two gray level threshold values are obtained, the histogram in the step 1) is cut by utilizing the threshold values to obtain three sub-histograms;
step 4), respectively calculating the probability density function and the cumulative distribution function of the two boundary sub-histograms;
step 5), respectively calculating transfer functions of the two boundary sub-histograms mapped to the middle sub-histogram by utilizing the cumulative distribution function;
step 6), pixels of the two boundary sub-histograms are redistributed to a middle sub-histogram area according to the two transfer functions, and an in-process image and a corresponding correction sub-histogram are obtained;
and 7) expanding the corrected sub-histograms to the global state, namely stretching the two-polarization gamma transformation of the histograms of the images in the process to obtain a final enhancement result.
2. The histogram transformation-based uniform exposure image contrast enhancement method according to claim 1, wherein: the method corrects the original histogram distribution based on histogram transformation, redistributes two-pole boundary pixels, and performs two-polarization adaptive gamma transformation stretching on the image according to the corrected histogram so as to obtain the image with enhanced contrast.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210664585.1A CN114998165A (en) | 2022-06-14 | 2022-06-14 | Histogram transformation based uniform exposure image contrast enhancement method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210664585.1A CN114998165A (en) | 2022-06-14 | 2022-06-14 | Histogram transformation based uniform exposure image contrast enhancement method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114998165A true CN114998165A (en) | 2022-09-02 |
Family
ID=83033848
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210664585.1A Pending CN114998165A (en) | 2022-06-14 | 2022-06-14 | Histogram transformation based uniform exposure image contrast enhancement method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114998165A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115797234A (en) * | 2023-01-29 | 2023-03-14 | 南京邮电大学 | Method for enhancing low-contrast two-dimensional code image recognition effect |
CN117409083A (en) * | 2023-12-14 | 2024-01-16 | 珠海市金锐电力科技有限公司 | Cable terminal identification method and device based on infrared image and improved YOLOV5 |
-
2022
- 2022-06-14 CN CN202210664585.1A patent/CN114998165A/en active Pending
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115797234A (en) * | 2023-01-29 | 2023-03-14 | 南京邮电大学 | Method for enhancing low-contrast two-dimensional code image recognition effect |
CN115797234B (en) * | 2023-01-29 | 2023-09-12 | 南京邮电大学 | Method for enhancing low-contrast two-dimensional code image recognition effect |
CN117409083A (en) * | 2023-12-14 | 2024-01-16 | 珠海市金锐电力科技有限公司 | Cable terminal identification method and device based on infrared image and improved YOLOV5 |
CN117409083B (en) * | 2023-12-14 | 2024-03-22 | 珠海市金锐电力科技有限公司 | Cable terminal identification method and device based on infrared image and improved YOLOV5 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108876743B (en) | Image rapid defogging method, system, terminal and storage medium | |
Yu et al. | Low-illumination image enhancement algorithm based on a physical lighting model | |
CN114998165A (en) | Histogram transformation based uniform exposure image contrast enhancement method | |
US7092573B2 (en) | Method and system for selectively applying enhancement to an image | |
CN108765336B (en) | Image defogging method based on dark and bright primary color prior and adaptive parameter optimization | |
Wang et al. | Automatic local exposure correction using bright channel prior for under-exposed images | |
KR20070075872A (en) | Image processing method and system using gain-controllable clipped histogram equalization | |
CN114331873B (en) | Non-uniform illumination color image correction method based on region division | |
CN111968041A (en) | Self-adaptive image enhancement method | |
CN111127343B (en) | Histogram double-control infrared image contrast enhancement method | |
US20130039577A1 (en) | Method for improving image quality | |
CN106709504B (en) | High fidelity tone mapping method for detail preservation | |
Yu et al. | Adaptive inverse hyperbolic tangent algorithm for dynamic contrast adjustment in displaying scenes | |
Subramani et al. | Optimal Bezier curve modification function for contrast degraded images | |
CN112419195A (en) | Image enhancement method based on nonlinear transformation | |
CN113947553B (en) | Image brightness enhancement method and device | |
CN109859138B (en) | Infrared image enhancement method based on human visual characteristics | |
CN108550124B (en) | Illumination compensation and image enhancement method based on bionic spiral | |
CN110728645A (en) | Image detail enhancement method and device based on guide filter regularization parameter and electronic equipment | |
CN105243651A (en) | Image edge enhancement method based on approximate variance and dark block pixel statistic information | |
CN117274085A (en) | Low-illumination image enhancement method and device | |
Chen et al. | An adaptive image dehazing algorithm based on dark channel prior | |
CN114283076A (en) | Image processing method based on transmissivity and CLAHE algorithm | |
CN110874822B (en) | Signal filtering method and system using dynamic window smoothing filter | |
CN114119433A (en) | Dark image processing method based on Bezier curve |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |